© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 127337
    Prev
    Next
    INDEX
    Explanations

    defluoridation, destock, demineralization

    np_acts-logits-general · gemini-2.5-flash-lite

    words related to removing or decreasing something, such as "denitration", "debiasing", "depressurize", and "demineralization".

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron activates on “de-” prefixed action words (like debiasing, denitration, depressurize, demineralization, destabilization, etc.)—i.e. verbs beginning with the “de-” prefix indicating removal or reversal.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     prized
    -2.17
     staunch
    -2.06
     begle
    -1.98
     Therefore
    -1.98
     hapless
    -1.94
     esteemed
    -1.89
     daunting
    -1.87
     Also
    -1.85
     erläut
    -1.85
     attain
    -1.84
    POSITIVE LOGITS
    芣
    2.25
     vokal
    2.23
     kupa
    2.23
     صفحۀ
    2.22
     mekan
    2.17
     optik
    2.17
     maaf
    2.13
     seguimos
    2.09
     tekno
    2.08
     素描
    2.08
    Activations Density 0.102%

    No Known Activations